Communication Network Management Data Mining Based on RPA and Parallel BP Network

In order to ensure the security of information transmission of enterprise Intranet and solve the problem of low accuracy of the traffic recognition method matched by traditional features, an recognition method based on improved BP and Spark cluster is proposed, and the method is applied to the traffic recognition of KDD99 data set and Robotic Process automation (RPA) traffic monitoring platform. The results show that compared with Artificial Neural Network (ANN) algorithm and Support Vector Machine (SVM) algorithm, the improved BP algorithm has higher recognition accuracy, accuracy, detection rate and false detection rate are 10.5%, 5.4% and 12.3%, 6.2%, respectively, and it shows good recognition performance. In the Spark cluster environment, the average training time and testing time of the improved Back Propagation (BP) neural network are 771.4s and 639.991s less than those of the improved BP model in single machine, respectively, showing that the recognition time is greatly shortened, and it has a good recognition rate. With the increase of cluster nodes, the training time of the improved BP model is always kept at a low level. It is concluded that the improved BP based on SPARK cluster+improved BP can be used to recognize abnormal traffic of Intranet transmission.